Hello,

as the title states: does anybody had expierence with running the operator environment as singel app preventing the operators running other applications besides the proview environment?
the goel would be starting up the operator station and bringing up the main xttgraph on one window and the alarm window on another screen.

Your kiosk application is the same as a stand-alone controller or an operator station, but don't try to do it with a workstation. From a workstation, distribute to your stand-alone kiosk controller.

On the kiosk machine, consider removing all of the Linux apps you don't want anyone to run, including browser, task bar and menus. Exactly how to do that depends on the flavor of Linux you are using. You should also create a startup script to automatically start your kiosk displays on a reboot. For example>
rt_xtt -qn Op &

In the Volume's OP object, you can specify one or more graphics to start, automatically. Proview already has start-up scripts built-in to restart a controller on a reboot. Most of your work is stripping down the Linux Kiosk machine to achieve the Spartan configuration you desire.



/ Ron

Thanks Ron,
the answer is a bit like I expected and when looking into the kiosk mode more and more I see that this is idd not the way to go. Since I'm used to Ubuntu and Debian those will probably one of the flavours I choose (in the past I've used a lot redhead, but since the pi came I moved over).

I'm going to try to start with a bare minimum installation and install gnome manually rather than removing apps one by one. Building an "idot proof" operator station is probably the most challeging job for the moment (I know my operators since I was once one of them).

Any tips and tricks are always more than welcome

Hi Jeroen,

Maybe to start with something like this?

Small Debian:
netinstall without Xorg

X11 basic:
apt-get install xserver-xorg-core xserver-xorg-video-all xserver-xorg-input-all x11-xserver-utils x11-xkb-utils x11-utils xinit xterm --no-install-recommends

Try: "startx" as user pwrp


If this works install ProviewR (have a fight with dependencies) and create the start-up scripts.

Create a "watchdog" every 5 seconds in the back ground and check if your operator station still is running.
If it's not running, kill all ProviewR processes and run the start-up script.

/Marc

I know my operators since I was once one of them

If you are willing to share, I am interested to know what type equipment and process you operated.

Is the vision for your kiosk a process control operator station with touch-screen? Or something like a bank ATM machine, fast food menu ordering system or airline ticketing booth?

I like Marc's suggestions. When building a stripped down system, I would lean towards a bare-bones display server, too.

For older versions of ProviewR, I have been using Lubuntu 16.04 with LXDE or xfce4 with good success. Recently, I have switched to Debian 12 LXDE to host Proview 6.1.1. A network of virtual Proview machines are hosted in a VMWare esxi server.

Wishing you good luck with your project. Keep us up-to-date on how it goes and please report your progress!

Ron

Hello,

The project I'm working on is a waste water managment and seal water control system. It will be used in the decommisioning phase of our power plant. The system controlled are a lot of existing pumps and 8 new pumps, there isn't a lot of logic involved, the most logic will be the operator faceplates and a system overview.
The initial plan was to run the whole system on 2 IPC's (also ESXI's) running debian 12 for each node. But since I'm going to reuse some PC's already installed with screens on the control desk, the operator stations will be on those existing PC's. The process nodes will be run in redundancy mode controlling about an 7 islands with a lot of IO's (mainly valve OLS/CLS, pump status and tank levels). So it will be a nice system .

In the meantime I've tried out Marc's solution and idd this is basic, but maybe a bit to basic so I'll try out some window managers. I see you mention LXDE, I'm going to give that oen a shot and I think there are a lot of resources available since the PI's are using it. Although some interresting tricks are always welcome (trying out not to reinvent the wheel here).
I must say the second reason a moved the operator stations towards an external PC was that I noticed some influence from my developer station on to my process nodes. it was eating resources and causing some delays in my PLC counts. Now I've installed the RT kernel in debian (reason why I switched from Ubuntu to Debian) and since then the performance is a lot better, but leaving me with some thinking about having my op stations onto the same esxi (the servers are Siemens IPC647D's, so not the newest ones).

So that's in short the story, a lot of work to do (hard and software), but I've still got some time: implementation will be end this year.